ClickAider
You are currently browsing the Bogle’s Blog weblog archives.

Adding an editorial voice to Google Search Results

In my post on Is Google Vulnerable in the search platform war, I incorrectly stated the following:

“Suppose you have compiled a list of the 20 best sources of reviews on the web, and you want to leverage Google’s full text index to search these. This isn’t possible to do, because Google allows only a single “site:” specified in a query.”

This, my friend, may have once been true, but it’s not true now. Google allows multiple site: specifiers to separated by an OR, and they have upped the number of terms allowed in a query to 32.

This enables people to layer an editorial voice on top of Google by specifying a set of sites that they think are authoritative on a particular topic. For example, in the sidebar to the right of this page, you can search the web pages of my friends and colleagues to find out what they think on any topic. View source on this page to see the simple Javascript that was used to combine the user query with site specifiers.

See what my colleagues know about marketing

A multi-site query is quite a useful feature and one that I think blog and web page authors will tap into widely once it becomes known. I can imagine tools that will automatically construct the appropriate query given a blogroll or favorites list.

Jobster V1 launched in San Diego!

One reasons my free time is something less than ample (in addition to two daughers) has been getting ready for the Jobster Launch, which took place on March 30 at the ERE Expo in San Diego.

It’s always incredibly busy before ship but I feel a great sense of excitement when something I’ve helped to create is about to reach the hands of users. Our launch event, which you can view here, including a live demo which was well received by the recruiters in attendance.

Jobster will be an invitation only service for professionals to begin with; this is part of our strategy of focusing on connecting companies with the best people rather than the most active job seekers. V1 is just the beginning, we’ve got many things in the works and there is much to come that will be of interest to every professional.

Side Projects for my ample free time

At any given time I have a set of fun little hacks I’m working on or wish I was working on. Here are the things I’m currently thinking about; in part I’m writing in the hope that someone will tell me that what I want already exists.

  • XBMC extensions for photo viewing and tagging: XBox media center is a great way to view computer content in the living room, and it’s open source and support for python scripting make it easy to extend. One extension I’d like to add is a way to make photo viewing interactive– I’d like to be able to hit a button to flag the best pictures for printing or tag them as a favorite. Another extension I’d like is Flickr for XBMC– a way to browse my Flickr collection on the Xbox. Given their web services API, this could probably be done entirely in Python.
  • Universal search plugin for Firefox: I like the search toolbar in Firefox, but it’s actually a bit of drag to have to select a search engine from the dropdown. Also, many sources of intranet content don’t yet have plugins, like Jobster’s corporate knowledge base in Confluence or our bug tracking database in JIRA. I’ve written a plugin that defers to a web page to intelligent select the appropriate search engine to based on the user’s query. If the search page doesn’t recognize the user’s query, it will default to Google, with a set of tabs across the top that allow you to run the same query on other search engines. If it recognizes the issue as a JIRA bug id, it will automatically go to that bug page in Jira. Special keywords allow the user to specify other search engines. All of the logic is in the web page, not the toolbar, so it’s easy to extend. Geeky but good.
  • Berry attachments: The Blackberry gives me instant, secure access to corporate email, but web pages inside the corporate firewall are off limits to me. I have also have no way to remotely search my Google Desktop. This program will be an Outlook plugin that recognizes specially formatted emails from me and replies back to me with the desired results, like the contents of the web page or the results of the Google Desktop search. This would be a bit like the Itrezzo, but hosted on your own PC, open source, extensible, and free.

    A potential security nightmare but I think I can make it safe enough for my purposes. The SpamBayes plugin, written in Python, will make a good starting point for this.

  • Berry Bloglines 2.0: Berry Bloglines has some unique content reformatting features, but many user’s have reported compatability problem with their Blackberries, probably having to do with the way cookies are handled. I’d like to fix this with a smart frontend to the browser based pages.
  • MSNBC - 290 dead reported in 8.7 quake near Indonesia

    MSNBC - 290 dead reported in 8.7 quake near Indonesia

    It’s sad to see another loss of life to an earthquake so soon after the last one. (I remember speculating when I was in Sri Lanka last January that seeing another quake of magnitude similar to the last one was unlikely.)

    In one sign of progress, it sounds like many countries had more effective evacauation warnings than the last time around.

    “Joel on Software” on Hiring

    Joel Spolsky writes about the challenges of using internet based advertising to find qualified applicants for Fog Creek Software.

    Until now we’ve been hiring rarely and quietly, but lately our sales are so strong we can’t quite keep up.

    My old theory of hiring was to post a job listing on Monster or Craigslist and then sort through the massive pile of unqualified applicants in hopes of finding the needle in the haystack.

    That hasn’t worked so well. In the future I’m going to try putting up semi-permanent job listings for all the kinds of people we might hire on the Fog Creek website and see if that gets us a slower trickle of more qualified job applicants.

    Spam is what happens whenever there is completely frictionless communication– witness email spam, resume spam, comment spam, search engine optimization spam, and more. It’s only a matter of time before social networks reach critical mass and start having relationship spam.

    A major focus of the next generation of internet services will be to create systems and communities that counter spam.

    Spam victims of all kinds are so busy fighting their current fires that it’s difficult to even consider alternatives. Alternatives often require more up front planning and more apparent effort. But at the end of the day the effort invested is far less than that spent fruitlessly wading through spam.

    Here Comes Location-Based Mobile Info

    Here Comes Location-Based Mobile Info

    About 10 days ago Nextel and MapQuest announced a service that delivers location-based maps and local info (shops, restaurants, etc.) to any Nextel GPS-enabled phone. The service is called MapQuest Find Me. The service also allows you to share your location with trusted people (via the web) and even send text messages through the service.

    flickr graph

    flickr graph - marcos weskamp

    Flickr Graph is an application that explores the social relationships inside flickr.com. It makes use of the classic attraction-repulsion algorithm for graphs. Start exploring your contacts by entering your flickr username or the email address you used to register there.

    Treo411, Pocket411?

    Berry 411 shouldn’t be limited to blackberries.

    Berry411 consists of a front end client that handles the initial stages of the user’s search, joined to backend web pages that display the search results. These web pages should work on any device that has an HTML browser.

    I don’t have time to reimplement the client on all the different smart phones out there, but I’d be happy to provide technical data on the URL formats expected by the backend for any who’d like to try.

    Please contact me if interested.

    Vertical Beat: Eliyon Becomes Zoominfo.com

    Vertical Beat: Eliyon Becomes Zoominfo.com

    Zoominfo.com is a people search service formerly known as Eliyon. The data is obtained from publically available sources such as press releases.

    The interface has been simplified and more of the service is available for free; everyone is hoping to follow in Google’s footsteps it seems.

    A lot of the people whom I most respect are not represented yet, even those with active blogs and a good web footprint. Worth watching since I’m sure it will improve over time.

    Robot Co-op: Towards an internet publicity policy

    Laurel Fan, a very talented developer, found these thoughts on “publicity polices” (as a counterpoint to “privacy policies”) on the Robot co-op web log.

    There’s an old New Yorker cartoon that shows some dogs at a computer with the caption “on the internet, no one knows you are a dog”. It’s a Web 1.0 joke about privacy and anonymity. But today, with the rise of blogs and websites like Flickr, del.icio.us and 43 Things, the joke is, “offline, no one knows you are a talented photographer, that you want to learn Italian, that you discover great links or know some great people”.

    For now, “The Real World” still requires you to check your virtual aspirations and accomplishments at the edge of the internet. You can’t convert your whuffie just yet. But millions of people are building online identities and reputations that are accumulating meaning and value. That value gets created as we build an online reputation, as we move our private knowledge into the public domain. The issues that this brings up get lost when the only policy folks have to reference is a “privacy policy”. Perhaps we need a “publicity policy” for 2.0 websites.

    Privacy policies focus on what happens to your private information. But a publicity policy could make clear what the implications are of sharing information in public.

    Search engines change the nature and significance of public information. They can gather and collate that information and make it conveniently accessible to complete strangers. For example there are search services that are capable of assembling a virtual resume of an individual by combining publically available information from across the web.

    What is the contract of trust between the searcher, the search engine, and the individual being searched?

    How can a search service allow individuals to opt in to the collation of publically available information about them, if the services has no involvement with that individual?

    Equally importantly, what recourse does the individual have when the search engine gathers information that presents a misleading or inaccurate impression?

    This is a very real possibility given the limitations of technology today.

    For example, the resume search service mentioned earlier incorrectly identified Linus Torvalds as a co-founder of Microsoft, as reported in Business Week. Linus is able to laugh of the inaccuracy, but for others the error could be serious.

    Enter “Microsoft (MSFT )” into the database and it spits out 5,200 names of “former” employees. Not so fast, though. The site lists Linus Torvalds as a “co-founder” of the software giant.

    That should come as news to anyone who follows technology and knows that Torvalds is the creator of the Linux operating system, perhaps the biggest threat to Microsoft’s Windows empire. But Torvalds, tongue firmly in cheek, says it’s all too true. “We’ve kept it quiet for years, since I left early on after some technical differences of opinion,” he writes in an e-mail.

    Torvalds goes on to explain how he was airbrushed out of an early company photo of co-founders Bill Gates and Paul Allen, and a handful of others. “Quite frankly, I think they should have airbrushed out the rest of the founders, too, if they really wanted it to look good, but hey, that may be just the bitterness talking.” Microsoft officials declined to comment.